There is an urgent need for predictive in vitro models to improve disease modeling and drug target identification and validation, especially for neurological disorders. Cerebral organoids, as alternative methods to in vivo studies, appear now as powerful tools to decipher complex biological processes thanks to their ability to recapitulate many features of the human brain. Combining these innovative models with microfluidic technologies, referred to as brain organoids-on-chips, allows us to model the microenvironment of several neuronal cell types in 3D. Thus, this platform opens new avenues to create a relevant in vitro approach for preclinical applications in neuroscience. The transfer to the pharmaceutical industry in drug discovery stages and the adoption of this approach by the scientific community requires the proposition of innovative microphysiological systems allowing the generation of reproducible cerebral organoids of high quality in terms of structural and functional maturation, and compatibility with automation processes and high-throughput screening. In this review, we will focus on the promising advantages of cerebral organoids for disease modeling and how their combination with microfluidic systems can enhance the reproducibility and quality of these in vitro models. Then, we will finish by explaining why brain organoids-on-chips could be considered promising platforms for pharmacological applications.

Cerebral organoids offer significant potential for neuroscience research as complex in vitro models that mimic human brain development. However, challenges related to their quality and reproducibility hinder their reliability. Discrepancies in morphology, size, cellular composition, and cytoarchitectural organization limit their applications, particularly in disease modeling, drug screening, and neurotoxicity testing. Critically, current methods for organoid characterization often lack standardization, restricting their broader applicability. To address the need for standardized quality assessment of cerebral organoids, we developed a Quality Control (QC) methodology for 60-day cortical organoids, evaluating five key criteria using a scoring system: morphology, size and growth profile, cellular composition, cytoarchitectural organization, and cytotoxicity. We implemented a hierarchical approach, beginning with non-invasive assessments to exclude low-quality organoids, while reserving in-depth analyses for those that passed the initial evaluation. To validate this framework, we exposed 60-day cortical organoids to graded doses of hydrogen peroxide (H ₂ O ₂ ), inducing a range of quality outcomes. The QC system demonstrated its robustness by accurately discriminating organoid qualities. Our proposed QC framework is designed to be user-friendly, flexible, and broadly applicable, making it suitable for routine assessment of cerebral organoid quality. Additionally, its scalability enables industrial applications, offering a valuable tool for advancing both fundamental and pre-clinical research.

In recent years, 3D cell culture systems have emerged as sophisticated in vitro models, providing valuable insights into human physiology and diseases. The transition from traditional 2D to advanced 3D cultures has introduced novel obstacles, complicating the characterization and analysis of these models. While the lactate dehydrogenase (LDH) activity assay has long been a standard readout for viability and cytotoxicity assessments in 2D cultures, its applicability in long-term 3D cultures is hindered by inappropriate normalization and low LDH stability over time. In response to these challenges, we propose an optimization of LDH assays, including a crucial normalization step based on total protein quantification and a storage method using an LDH preservation buffer. We applied it to compare unexposed cerebral organoids with organoids exposed to a toxic dose of valproic acid, and showed efficient normalization of cellular viability as well as enhanced LDH stability within the buffer. Importantly, normalized LDH activity results obtained were independent of organoid dimension and cell density. This refined LDH assay, tailored to address 3D culture constraints, allows for the transposition of this routine test from 2D to 3D cultures.

Cerebral organoids hold great promise for neuroscience research as complex in vitro models that mimic human brain development. However, they face significant challenges related to quality and reproducibility, leading to unreliability in both academic and industrial contexts. Discrepancies in morphology, size, cellular composition, and cytoarchitectural organization limit their application in biomedical studies, particularly in disease modeling, drug screening, and neurotoxicity testing, where consistent models are essential. Critically, current methods for organoid characterization often lack standardization and rely heavily on subjective assessments, restricting their broader applicability. In this study, we developed a comprehensive Quality Control (QC) framework for 60-days cortical organoids. Five key criteria: morphology, size and growth profile, cellular composition, cytoarchitectural organization, and cytotoxicity, are evaluated using a standardized scoring system. We implemented a hierarchical approach, beginning with non-invasive assessments to exclude low-quality organoids (Initial Scoring), while reserving in-depth analyses for those that passed the initial evaluation (Final Scoring). To validate this framework, we exposed 60-day cortical organoids to graded doses of hydrogen peroxide (H ₂ O ₂ ), inducing a spectrum of quality outcomes. The QC system demonstrated its robustness and reproducibility by accurately discriminating organoid quality based on objective and quantifiable metrics. This standardized and user-friendly framework for quality assessment not only minimizes observer bias but also enhances the reliability and comparability of cerebral organoid studies. Additionally, its scalability makes it suitable for industrial applications and adaptable to other organoid types, offering a valuable tool for advancing both fundamental and preclinical research.

Cerebral organoids have emerged as advanced three-dimensional in vitro models capable of recapitulating essential aspects of human brain development, thereby offering new opportunities for studying brain physiology and disease. Nonetheless, their use in preclinical settings remains limited by their high variability. Microfluidic devices offer a promising alternative to conventional culture supports by enabling improved recapitulation of the cellular microenvironment; however, most cerebral organoids-on-chips rely on custom-made academic prototypes, limiting their scalability and translatability. Here, we established a standardized brain organoid-on-chip model by combining a regionalized cortical organoid differentiation protocol with a commercially available microfluidic device (Duplex™, NETRI). Cortical organoid quality and reproducibility were evaluated using a previously established multi-criteria Quality Control (QC) framework based on a scoring system. To enhance objectivity and throughput, the non-invasive morphometric QC criteria were automated using deep learning-based image analysis. Using a single human induced pluripotent stem cell (hiPSC) line, we assessed whether microfluidic culture conditions could improve the quality and reproducibility of cortical organoids within a controlled genetic background. Compared with conventional cultures, on-chip cortical organoids displayed improved inter- and intra-batch reproducibility across multiple scales, including morphology, size, cytoarchitectural organization, and transcript expression, together with higher QC scores and reduced QC failure rates. Finally, the QC framework was adapted for neurotoxicity assessment using a decision tree to classify compounds according to their neurotoxic risk. While this study represents a proof-of-concept demonstration within a single genetic background, the integrated brain organoid-on-chip platform and standardized QC evaluations provide a flexible and translatable approach for preclinical and industrial studies and support future validation across multiple hiPSC lines.